Visualizations
Adeline Shin
4/26/2020
Load Data
load("./curves.RData")Plotting Original Data
real_df = by_country %>%
ungroup(country_region) %>%
mutate(
group = replicate(length(by_country$country_region), 0)
) %>%
group_by(country_region) %>%
mutate(
log_cases = log10(confirmed_cases),
max_log = max(log_cases)
) %>%
dplyr::select(-log_cases) %>%
arrange(max_log)
for (i in 1:length(real_df$country_region)) {
if (real_df$max_log[i] < 2) {
real_df$group[i] = "0 - 99"
} else if (real_df$max_log[i] < log10(500)) {
real_df$group[i] = "100 - 499"
} else if (real_df$max_log[i] < 3) {
real_df$group[i] = "500 - 999"
} else if (real_df$max_log[i] < log10(5000)) {
real_df$group[i] = "1000 - 4999"
} else if (real_df$max_log[i] < 4) {
real_df$group[i] = "5000 - 9999"
} else if (real_df$max_log[i] < 5) {
real_df$group[i] = "10000 - 99999"
} else {
real_df$group[i] = "100000 +"
}
}
real_df$group = factor(real_df$group, levels = c("0 - 99", "100 - 499", "500 - 999", "1000 - 4999", "5000 - 9999", "10000 - 99999", "100000 +"), ordered = TRUE)
real_df %>%
ggplot(aes(x = t, y = confirmed_cases)) +
geom_path(aes(color = country_region)) +
facet_grid(group ~ ., scales = "free") +
theme(legend.position = "none") +
geom_dl(aes(label = country_region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for All Countries (Real Data)",
x = "Days Since First Case",
y = "Cumulative Cases (Grouped by Total)"
)
ggsave("./visualizations/real_data_plotted.jpg", width = 12, height = 8)Plotting Estimates
Plotting Top 20 Countries with Greatest Population
fitted_list_2 = as.data.frame(unlist(fitted_list)) %>%
janitor::clean_names() %>%
rownames_to_column(var = "country") %>%
mutate(
t = as.numeric(gsub("[^[:digit:]]", "", country)) - 1,
region = gsub("[[:digit:]]","", country)) %>%
dplyr::select(region, t, cases = unlist_fitted_list)
large_pop_list = c("China", "India", "US", "Indonesia", "Pakistan", "Brazil", "Nigeria", "Bangladesh", "Russia", "Mexico", "Japan", "Ethiopia", "Phillipines", "Egypt", "Vietnam", "Congo (Kinshasa)", "Turkey", "Iran", "Germany", "Thailand")
top_population = fitted_list_2 %>%
filter(region %in% large_pop_list) %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for Top 20 Countries in Population",
x = "Days After First Case",
y = "Cumulative Cases"
)
ggsave("./visualizations/top_20_regions.jpg", plot = top_population, width = 12, height = 8)Comparison of Real and Estimated Data
comparison_df = by_country %>%
rename(region = country_region) %>%
dplyr::select(-fatalities) %>%
left_join(fitted_list_2, by = c("region", "t")) %>%
rename(estimated_cases = cases) %>%
filter(region == "US" | region == "China" | region == "Italy" | region == "Iran") %>%
pivot_longer(confirmed_cases:estimated_cases, names_to = "data_type")
comparison_df %>%
ggplot(aes(x = t, y = value)) +
geom_path(aes(color = data_type)) +
facet_wrap(region ~ ., scale = "free", nrow = 2) +
labs(
title = "Real Data vs. Estimated Data from Algorithm for Select Countries",
x = "Days Since First Case",
y = "Number of Cumulative Cases"
)
ggsave("./visualizations/comparison_graph.jpg", width = 12, height = 8)Plotting Top 20 Countries with Most Cases
top_names = fitted_list_2 %>%
group_by(region) %>%
mutate(max_cases = max(cases)) %>%
dplyr::select(region, max_cases) %>%
distinct() %>%
arrange(desc(max_cases)) %>%
head(20) %>%
dplyr::select(region)
top_names = as.tibble(top_names)
top_cases = fitted_list_2 %>%
filter(region %in% top_names$region) %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for Countries with Most Cases",
x = "Days After First Case",
y = "Cumulative Cases"
)
ggsave("./visualizations/highest_cases.jpg", plot = top_cases, width = 12, height = 8)Plotting Countries Where Total is Still Growing
growing_df = fitted_list_2 %>%
full_join(param_df1, by = "region")
for (i in 1:length(growing_df$region)) {
if (growing_df$b[i] < 0.08) {
growing_df$group[i] = "Growing Very Slowly"
} else if (growing_df$b[i] < 0.15) {
growing_df$group[i] = "Growing Slowly"
} else if (growing_df$b[i] < 0.22) {
growing_df$group[i] = "Growing Moderately"
} else if (growing_df$b[i] < 0.32) {
growing_df$group[i] = "Growing Quickly"
} else {
growing_df$group[i] = "Growing Very Quickly"
}
}
growing_df$group = factor(growing_df$group, levels = c("Growing Very Slowly", "Growing Slowly", "Growing Moderately", "Growing Quickly", "Growing Very Quickly"), ordered = TRUE)
growing_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
facet_grid(group ~ ., scales = "free") +
theme(legend.position = "none") +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for All Countries",
x = "Days Since First Case",
y = "Cumulative Cases (Grouped by Growth Rate)"
)
ggsave("./visualizations/growth_group_plot.jpg", width = 12, height = 8)Plotting Countries by Continent
continent_df =
fitted_list_2 %>%
mutate(
continent = countrycode(sourcevar = fitted_list_2[, "region"],
origin = "country.name",
destination = "continent")
)
continent_df$continent[4320:4345] = "Europe"
continent_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
facet_grid(continent ~ ., scales = "free") +
theme(legend.position = "none") +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for All Countries",
x = "Days Since First Case",
y = "Cumulative Cases (Grouped by Continent)"
)
ggsave("./visualizations/covid_19_continents.jpg", width = 12, height = 8)Plotting Countries by Number of Cases
number_df = fitted_list_2 %>%
mutate(
group = replicate(length(fitted_list_2$cases), 0)
) %>%
group_by(region) %>%
mutate(
log_cases = log10(cases),
max_log = max(log_cases)
) %>%
dplyr::select(-log_cases) %>%
arrange(max_log)
for (i in 1:length(number_df$region)) {
if (number_df$max_log[i] < 2) {
number_df$group[i] = "0 - 99"
} else if (number_df$max_log[i] < log10(500)) {
number_df$group[i] = "100 - 499"
} else if (number_df$max_log[i] < 3) {
number_df$group[i] = "500 - 999"
} else if (number_df$max_log[i] < log10(5000)) {
number_df$group[i] = "1000 - 4999"
} else if (number_df$max_log[i] < 4) {
number_df$group[i] = "5000 - 9999"
} else if (number_df$max_log[i] < 5) {
number_df$group[i] = "10000 - 99999"
} else {
number_df$group[i] = "100000 +"
}
}
number_df$group = factor(number_df$group, levels = c("0 - 99", "100 - 499", "500 - 999", "1000 - 4999", "5000 - 9999", "10000 - 99999", "100000 +"), ordered = TRUE)
number_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
facet_grid(group ~ ., scales = "free") +
theme(legend.position = "none") +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases for All Countries",
x = "Days Since First Case",
y = "Cumulative Cases (Grouped by Total)"
)
ggsave("./visualizations/all_covid_cases.jpg", width = 12, height = 8)Differences in Stay at Home Orders
distancing_df = fitted_list_2 %>%
filter(region == "US" | region == "China" | region == "Italy" | region == "Sweden") %>%
mutate(population = rep(0, length(cases)))
for (i in 1:length(distancing_df$cases)) {
if (distancing_df$region[i] == "US") {
distancing_df$population[i] = 328200000
} else if (distancing_df$region[i] == "China") {
distancing_df$population[i] = 1393000000
} else if (distancing_df$region[i] == "Italy" ) {
distancing_df$population[i] = 60360000
} else {distancing_df$population[i] = 10230000}
}
distancing_df = distancing_df %>%
mutate(
percent_infected = (cases / population) * 100
)
highlights = unlist(x = c(distancing_df[14, ], distancing_df[286, ], distancing_df[214, ]))
distancing_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
geom_point(data = distancing_df[14,], color = "#F8766D") +
geom_point(data = distancing_df[286,], color = "#C77CFF") +
geom_point(data = distancing_df[130,], color = "#7CAE00") +
labs(
title = "Estimated Curves for Select Countries and Stay At Home Implementation Date",
x = "Days Since First Case (Black Dot Indicates Implementation of Social Distancing)",
y = "Cumulative Cases"
)
ggsave("./visualizations/social_distancing.jpg", width = 12, height = 8)
distancing_df %>%
ggplot(aes(x = t, y = percent_infected)) +
geom_path(aes(color = region)) +
geom_point(data = distancing_df[14,], color = "#F8766D") +
geom_point(data = distancing_df[286,], color = "#C77CFF") +
geom_point(data = distancing_df[130,], color = "#7CAE00") +
labs(
title = "Estimated Curves for Select Countries and Stay At Home Implementation Date",
x = "Days Since First Case (Dot Indicates Implementation of Social Distancing)",
y = "Cumulative Percentage Infected"
)
ggsave("./visualizations/social_distancing_percentage.jpg", width = 12, height = 8)Plotting Clustering Results
K means Clustering
kmeans_df = read_csv("./param_km3_final.csv") %>%
group_by(cluster) %>%
mutate(
mean_a = round(mean(a), digits = 0),
mean_b = round(mean(b), digits = 2),
mean_c = round(mean(c), digits = 2)
)## Parsed with column specification:
## cols(
## X1 = col_double(),
## a_std = col_double(),
## b_std = col_double(),
## c_std = col_double(),
## region = col_character(),
## a = col_double(),
## b = col_double(),
## c = col_double(),
## cluster = col_double()
## )kmeans_graphing_df = fitted_list_2 %>%
left_join(kmeans_df, by = "region") %>%
dplyr::select(region, t, cases, cluster)
labels = c(`1` = "Cluster 1: a = 198895, b = 0.26, c = 39.61", `2` = "Cluster 2: a = 24196, b = 0.08, c = 73.55", `3` = "Cluster 3: a = 7970, b = 0.13, c = 35.23")
kmeans_graphing_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
facet_grid(cluster ~ ., scales = "free", labeller = labeller(cluster = labels)) +
theme(legend.position = "none") +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases Grouped by K-Means Clustering",
x = "Days Since First Case",
y = "Cumulative Cases Grouped by Cluster"
)
ggsave("./visualizations/kmeans_graph.jpg", width = 12, height = 12)kmeans_plot = kmeans_graphing_df %>%
ggplot(aes(x = t, y = cases, color = cluster, text = region)) +
geom_path() +
theme(legend.position = "none") +
labs(
title = "Cumulative COVID-19 Cases Grouped by K-Means Clustering",
x = "Days Since First Case",
y = "Cumulative Cases Grouped by Cluster"
)
ggplotly(kmeans_plot, tooltip = "text")Gaussian Clustering
gmm_df = read_csv("./param_gmm_final.csv") %>%
group_by(cluster) %>%
mutate(
mean_a = round(mean(a), digits = 0),
mean_b = round(mean(b), digits = 2),
mean_c = round(mean(c), digits = 2)
)## Parsed with column specification:
## cols(
## X1 = col_double(),
## a_std = col_double(),
## b_std = col_double(),
## c_std = col_double(),
## region = col_character(),
## a = col_double(),
## b = col_double(),
## c = col_double(),
## cluster = col_double()
## )gmm_graphing_df = fitted_list_2 %>%
left_join(gmm_df, by = "region") %>%
dplyr::select(region, t, cases, cluster)
labels_g = c(`1` = "Cluster 1: a = 3983, b = 0.09, c = 55.17", `2` = "Cluster 2: a = 326117, b = 0.16, c = 65.44", `3` = "Cluster 3: a = 11998, b = 0.2, c = 36.34")
gmm_graphing_df %>%
ggplot(aes(x = t, y = cases)) +
geom_path(aes(color = region)) +
facet_grid(cluster ~ ., scales = "free", labeller = labeller(cluster = labels_g)) +
theme(legend.position = "none") +
geom_dl(aes(label = region), method = list(dl.combine("last.points"), cex = 0.7)) +
labs(
title = "Cumulative COVID-19 Cases Grouped by Gaussian Mixture Model Clustering",
x = "Days Since First Case",
y = "Cumulative Cases Grouped by Cluster"
)
ggsave("./visualizations/gmm_graph.jpg", width = 12, height = 12)gmm_plot = gmm_graphing_df %>%
ggplot(aes(x = t, y = cases, color = cluster, text = region)) +
geom_path() +
theme(legend.position = "none") +
labs(
title = "Cumulative COVID-19 Cases Grouped by Gaussian Mixture Model Clustering",
x = "Days Since First Case",
y = "Cumulative Cases Grouped by Cluster"
)
ggplotly(gmm_plot, tooltip = "text")income_group_df = read_csv("./country_code_gdp_indicators.csv") %>%
janitor::clean_names() %>%
rename("region" = table_name,
"global_region" = region) %>%
mutate(region = recode(region,
"Bahamas, The" = "Bahamas",
"Brunei Darussalam" = "Brunei",
"Côte d'Ivoire" = "Cote d'Ivoire",
"Congo, Dem. Rep." = "Congo (Kinshasa)",
"Congo, Rep." = "Congo (Brazzaville)",
"Czech Republic" = "Czechia",
"Egypt, Arab Rep." = "Egypt",
"Gambia, The" = "Gambia",
"Iran, Islamic Rep." = "Iran",
"Korea, Rep." = "Korea, South",
"Kyrgyz Republic" = "Kyrgyzstan",
"Lao PDR" = "Laos",
"Russian Federation" = "Russia",
"Myanmar" = "Burma",
"St. Kitts and Nevis" = "Saint Kitts and Nevis",
"St. Vincent and the Grenadines" = "Saint Vincent and the Grenadines",
"St. Lucia" = "Saint Lucia",
"São Tomé and Principe" = "Sao Tome and Principe",
"Slovak Republic" = "Slovakia",
"Syrian Arab Republic" = "Syria",
"United States" = "US",
"Venezuela, RB" = "Venezuela",
"Yemen, Rep." = "Yemen"))## Parsed with column specification:
## cols(
## `Country Code` = col_character(),
## Region = col_character(),
## IncomeGroup = col_character(),
## SpecialNotes = col_character(),
## TableName = col_character()
## )global_bank_df = kmeans_graphing_df %>%
left_join(income_group_df, by = "region") %>%
filter(!is.na(income_group))
income_group_plot = global_bank_df %>%
ggplot(aes(x = t, y = cases, color = cluster, text = region)) +
geom_path() +
facet_grid(income_group ~ ., scales = "free", labeller = labeller(cluster = labels)) +
theme(legend.position = "none") +
labs(
title = "Cumulative COVID-19 Cases Grouped by K-Means Clustering",
x = "Days Since First Case",
y = "Cumulative Cases Grouped by Cluster"
)
ggplotly(income_group_plot, tooltip = "text")